politicalbetting.com » Blog Archive » Predicting the election: What the leading academic teams ar

TheScreamingEagles said:

Wankfest?

Honest I didn't write that.

Anorak said:

TheScreamingEagles said:

Wankfest?

Honest I didn't write that.

Also, one word or two? Pick the answer and stick with it.

TheScreamingEagles said:

Wankfest?

Honest I didn't write that.

Richard_Nabavi said:

So, when you get 12 groups of academics, every one of which is forecasting a hung parliament, what should you do?

bigjohnowls said:

TheScreamingEagles said:

Wankfest?

Honest I didn't write that.

Were you with the 2nd forecaster!!!!

Richard_Nabavi said:

So, when you get 12 groups of academics, every one of which is forecasting a hung parliament, what should you do?

Richard_Nabavi said:

So, when you get 12 groups of academics, every one of which is forecasting a hung parliament, what should you do?

Roger said:

ARSE is so far out from all the other predictions do we put it down as another PB embarrassment like Angus Reid or merely an outlier?

antifrank said:

Richard_Nabavi said:

So, when you get 12 groups of academics, every one of which is forecasting a hung parliament, what should you do?

Consider the alternative, of course.

dr_spyn said:

http://www.itv.com/news/update/2015-03-27/first-reporter-convicted-in-elveden-trial-has-conviction-quashed/

Pulpstar said:

Casino_Royale said:

On 12th Feb, at the London polling event I went to, Lebo forecast Con - 322 and Lab - 254.

Have Dave's ratings tanked that much since then ?!

Peter_the_Punter said:

Richard_Nabavi said:

So, when you get 12 groups of academics, every one of which is forecasting a hung parliament, what should you do?

What's more, Richard, none of them puts either Party within 19 seats of an overall majority (and rather perversely Labour come closest with 306).

You can't often say that a 1/5 shot is value, but NOM at that price with Betfair is an exception.

Sunil_Prasannan said:

How come a lot of the predictions have stuck LDs and SNP or both in with "others"?

Pulpstar said:

OblitusSumMe said:

antifrank said:

Richard_Nabavi said:

So, when you get 12 groups of academics, every one of which is forecasting a hung parliament, what should you do?

Consider the alternative, of course.

Labour would need about a hundred gains from the Conservatives to win a majority at the same time as losing Scotland to the SNP, so the only realistic alternative is a Conservative majority - I think Pulpstar and others already came to the conclusion this morning...

Yes, I've backed Con majority for £40 at 7.8 and now it's drifted to 8.2/8.6 !

Casino_Royale said:

S&W model forecasts the best part of 85 Labour gains from the Tories in E&W.

I'd like to have some of what they're smoking.

Sean_F said:

Casino_Royale said:

S&W model forecasts the best part of 85 Labour gains from the Tories in E&W.

I'd like to have some of what they're smoking.

The Conservatives would have to be losing seats like Basildon South and East Thurrock, Burton, Reading West, Rugby, Battersea. For that to happen, they'd have to be 2-3% behind in England and Wales. Not completely beyond the bounds of possibility, if it's a really bad night, but significantly worse than current polling.

surbiton said:

Easterross said:

Good afternoon all and what struck me was the substantial change in Stephen Fisher's team prediction from last week. Have read their reasoning beyond the methodology which whether correct or not, is easy to follow.

This brings me back to my recurring comment, what chance Labour end up with fewer seats on 8th May than in 2010?

Zero

surbiton said:

Easterross said:

Good afternoon all and what struck me was the substantial change in Stephen Fisher's team prediction from last week. Have read their reasoning beyond the methodology which whether correct or not, is easy to follow.

This brings me back to my recurring comment, what chance Labour end up with fewer seats on 8th May than in 2010?

Zero

NickPalmer said:

Candidates are all getting floods of NGO material. A small prize for drama goes to this one:
-----
UK's FUTURE ON THE LINE

With a controversial referendum in Scotland behind us – for the time being - a General Election and a new Government just around the corner, increasing devolution of power around the country, and the longer-term prospect of a referendum and further renegotiation of the UK’s relationship with Europe, this is a critical time for the UK.

Prospective parliamentary candidates will be honing their messages to voters, and, if past turn-out is an indicator of future engagement in General Elections, those carefully-thought-out words will fall on largely deaf ears – so it’s time to talk about some of the issues that affect real people every day…
-----
So what is it about? Er, it's a proposal to paint road markings more often, from the Road Safety markings Association.

Richard_Nabavi said:

antifrank said:

One of the things that we are seeing in this table is the problem of seeing everyone's best guess. This means that the results cluster much more closely than they should. Is it unthinkable that Labour will get fewer than 260 seats or more than 315? Of course it isn't - that's quite a narrow band. Similarly, is it unthinkable that UKIP will get more than five seats? No, of course not.

I expect most of these forecasters would readily acknowledge the possibility of other results.

Probability distributions would be more helpful.

Fisher of course does provide error bars, currently for example he has:

Con: 296 (251– 343 at 95% confidence limits)

So, yes, in general commentators and punters under-estimate the probability of reasonably large shifts from the current position. Of course as the election gets closer polls start to become more and more reliable as predictors. We're pretty close now, so I wouldn't expect big shifts from current polling - however a shift of a few points makes quite a big difference in seats, so the full outcome is quite uncertain.

Overall, it looks like a hung parliament, Con most seats slightly more likely than Lab, Lab Maj out of reach, Con Maj still in reach but unlikely, SNP surge nailed on, LibDems floundering, UKIP around 3.

Basically everyone loses apart from the SNP.

antifrank said:

Richard_Nabavi said:

antifrank said:

One of the things that we are seeing in this table is the problem of seeing everyone's best guess. This means that the results cluster much more closely than they should. Is it unthinkable that Labour will get fewer than 260 seats or more than 315? Of course it isn't - that's quite a narrow band. Similarly, is it unthinkable that UKIP will get more than five seats? No, of course not.

I expect most of these forecasters would readily acknowledge the possibility of other results.

Probability distributions would be more helpful.

Fisher of course does provide error bars, currently for example he has:

Con: 296 (251– 343 at 95% confidence limits)

So, yes, in general commentators and punters under-estimate the probability of reasonably large shifts from the current position. Of course as the election gets closer polls start to become more and more reliable as predictors. We're pretty close now, so I wouldn't expect big shifts from current polling - however a shift of a few points makes quite a big difference in seats, so the full outcome is quite uncertain.

Overall, it looks like a hung parliament, Con most seats slightly more likely than Lab, Lab Maj out of reach, Con Maj still in reach but unlikely, SNP surge nailed on, LibDems floundering, UKIP around 3.

Basically everyone loses apart from the SNP.

I would add in the wildcard that isam keeps noting, which is that polling this time round is not necessarily as reliable as in previous cycles because the pollsters don't know what to make of the kippers (and to a lesser extent the SNP). So while opinion may be slowly setting, we can't be too confident that the pollsters are accurately picking up how it is setting.

isam said:

antifrank said:

Richard_Nabavi said:

antifrank said:

One of the things that we are seeing in this table is the problem of seeing everyone's best guess. This means that the results cluster much more closely than they should. Is it unthinkable that Labour will get fewer than 260 seats or more than 315? Of course it isn't - that's quite a narrow band. Similarly, is it unthinkable that UKIP will get more than five seats? No, of course not.

I expect most of these forecasters would readily acknowledge the possibility of other results.

Probability distributions would be more helpful.

Fisher of course does provide error bars, currently for example he has:

Con: 296 (251– 343 at 95% confidence limits)

So, yes, in general commentators and punters under-estimate the probability of reasonably large shifts from the current position. Of course as the election gets closer polls start to become more and more reliable as predictors. We're pretty close now, so I wouldn't expect big shifts from current polling - however a shift of a few points makes quite a big difference in seats, so the full outcome is quite uncertain.

Overall, it looks like a hung parliament, Con most seats slightly more likely than Lab, Lab Maj out of reach, Con Maj still in reach but unlikely, SNP surge nailed on, LibDems floundering, UKIP around 3.

Basically everyone loses apart from the SNP.

I would add in the wildcard that isam keeps noting, which is that polling this time round is not necessarily as reliable as in previous cycles because the pollsters don't know what to make of the kippers (and to a lesser extent the SNP). So while opinion may be slowly setting, we can't be too confident that the pollsters are accurately picking up how it is setting.

Thanks for the name check!

What I find quite mystifying are the Ukip scores in the last three ICM polls

One UK (or GB I don't know the difference!)
One Scotland only
One London only

Ukip got 9%, 7% & 9%

I can't make sense of that.

bigjohnowls said:

One forecaster thinks 9 SNP

Scott_P said:

It was a significant moment, because it was the Syria vote that Miliband himself selected as the best example of his suitability to be Prime Minister of this country.

And it was a lie. A total, and utter misrepresentation of what Ed Miliband did, and did not do, over the Syria vote. He knows it’s a lie, the shadow cabinet know it’s a lie, Labour MPs know it’s a lie.

Pulpstar said:

Sean_F said:

Casino_Royale said:

S&W model forecasts the best part of 85 Labour gains from the Tories in E&W.

I'd like to have some of what they're smoking.

The Conservatives would have to be losing seats like Basildon South and East Thurrock, Burton, Reading West, Rugby, Battersea. For that to happen, they'd have to be 2-3% behind in England and Wales. Not completely beyond the bounds of possibility, if it's a really bad night, but significantly worse than current polling.

I've got Reading West down as an outside shot for Labour